Reconnaissance geology of Lake McRae, Marlborough: Landslides, lake sediments and paleo-earthquakes

Abstract

Lake McRae is a small alpine (0.65 km2, 900 m a.s.l.) landslide-dammed lake nestled within the Inland Kaikoura Ranges in Marlborough. Results from cosmogenic dating indicate that the lake formed near the end of the last ice age when two very large landslides (LS1, LS2) blocked an east-west fault-parallel valley fed mainly from Goat Valley Stream. We conducted a reconnaissance investigation of the lake and its environs, looking at the central section of the Clarence Fault adjacent to the lake, the morphology and age of the landslides, and the lake bathymetry and stratigraphy.

The Clarence Fault was mapped within the study area. The fault cuts both landslides, crosses the floor of Lake McRae and traverses a bedrock slope above the lake. A hand-dug trench into deposits at the toe of a scarp along this shutter ridge yielded colluvial deposits that are probably related to surface faulting events. Based on two radiocarbon dates, these deposits are suggestive of 1 or perhaps 2 faulting events during the last 1900-2000 years. We estimate a long term slip rate for the central section of the Clarence Fault as c. 2.8 ± 1.4 mm/yr.

Bathymetric mapping of the lake floor shows that at its western end part of landslide LS1 is submerged and draped with sediment. Shallow seismic sections indicate a layered sedimentary fill and also help to identify the Clarence Fault within the lake. Shallow (1-m) cores taken from the deepest parts of the lake penetrate the upper part of the lake section. Three radiocarbon dates help define the age structure of the uppermost lake section. Calculation of sediment accumulation rates indicates that the lake has been filled for at least 10,000 years, which is consistent with the cosmogenic ages obtained from landslide LS1.

No slump or turbidite deposits were identified in the 1-m lake cores, which suggests that there has been no shaking of Modified Mercalli (MM) intensity IX during the last 1000 years. This means that it is unlikely that the nearby Clarence or Elliott faults have ruptured during this period. Increased sediment flux into the lake has been recognised in the uppermost part of the shallow lake cores. This phenomenon is more likely related to historic farming practice (burning, clearance) and rabbit infestation during the mid-19th century, rather than being linked to a specific earthquake event such as the 1848 M 7.5 Marlborough earthquake.

Our results indicate Lake McRae formed at the end of the last ice age (15.2 [+3.8/-2.7] thousand years ago) when two very large landslides blocked an east-west fault parallel valley. The lake has since been accumulating sediments derived mainly from Goat Valley Stream. The preliminary record from Lake McRae indicates that it has significant potential for further integrated landslide, paleoearthquake, ground motion, and lacustrine studies. In light of the recent November 14, 2016 Mw7.8 Kaikoura earthquake sequence, further work on the active faulting and sedimentary record of this important small basin is vital.

---

Technical Abstract

The Marlborough region is one of the most seismically and tectonically active parts of New Zealand as evidenced by the recent occurrence of the Mw 7.8 November 14, 2016 Kaikoura earthquake sequence. The Marlborough Fault System (MFS) is traversed by numerous active faults capable of producing large earthquakes Despite this little is known about its paleoseismic and geomorphic history. Lake McRae is the largest alpine lake in Marlborough and a study of its bottom sediments offered an opportunity to investigate the pre-European seismic and geomorphic history of this part of New Zealand. We undertook a reconnaissance of Lake McRae and its environs including: an investigation of the central section of the Clarence Fault; a study of the landslides that bound the lake; and, a bathymetric survey and sediment collection from the lake bed to determine its depth, age and stratigraphic characteristics.

Lake McRae is a small alpine (0.65 km2, 900 m a.s.l.) landslide-dammed lake nestled within the Inland Kaikoura Ranges. The lake formed near the end of the last ice age when two very large landslides blocked an east-west fault-parallel valley fed mainly from the northeast by Goat Valley Stream. This project involved a reconnaissance investigation of the lake and its environs, looking at the central section of the Clarence Fault adjacent to the lake, the morphology and age of the very large landslides, and of the bathymetry, stratigraphy and chronology of deposits in the lake.

Landslide LS1 (0.5 km2 and c. 41 x 106 m3), at the western end of the lake came from the north side of the valley that parallels the Clarence Fault. Cosmogenic 10Be ages of three greywacke tors and boulders protruding from the top of LS1 yield a mean age of 15,200 [+3800/-2700] years. Based on its morphology the larger, eastern (LS2; 0.5 km2 and c. 70 x 106 m3) landslide is probably somewhat older than LS1. LS2 was came from the southern range and collapsed to the north toward Goat Valley Stream.

The Clarence Fault was mapped within the study area. A shallow (0.7 m) trench was hand- dug into deposits at the toe of a scarp along the fault. The 2 m long trench exposed colluvial deposits derived from the slope of the shutter scarp. These deposits are mostly considered to relate to surface-faulting events that shed debris from the scarp. Based on a single radiocarbon date, these deposits indicate 1 or perhaps 2 faulting events during the last 2000 years. However, these results are unproven because fault strands were not exposed in the trench. Long-term slip rate estimates for the central section of the Clarence Fault and Elliott Fault were derived by reconstructing the history of the Acheron River, which traverses the Clarence-Elliott fault wedge from north to south. Their dextral fault slip rates are estimated at c. 2.8 ± 1.4 mm/yr and 1.4 ± 0.7 mm/yr, respectively, averaged through the Quaternary period.

The lake has a maximum water depth of c. 43 m. The bathymetry indicates the hummocky morphology of part of landslide LS1 on the floor of the lake at its western end. Sub-bottom CHIRP (shallow geophysical) sections indicate a layered sedimentary lake fill with an estimated thickness of up to 10 m. The CHIRP sections also help to identify the Clarence Fault in the subsurface in the southwest corner of the lake. Shallow (1-m) cores were taken from the deepest parts of the lake and penetrate the upper part of the lake section. The lake fill is comprised of fine-grained layered sedimentary packages. A basal radiocarbon date on plant fragments from these cores (at a depth of 0.64 m) is 416 ± 22 yr B.P. Projection of the rates of recent sediment accumulation implies that the lake has been filled (and accumulating layered lake sediments) for at least 10,000 years. This age is consistent with the cosmogenic 10Be ages from landslide LS1.
No lacustrine mass-wasting deposits were identified in the 1-m cores, which suggests that there has been no shaking of Modified Mercalli (MM) intensity IX during the last 1000 years. This means that it is unlikely that the Clarence or Elliott faults have ruptured during this time period. Increased sediment flux into the lake has been recognised in the uppermost part of the shallow lake cores. This phenomenon is believed to be related to recent strong shaking (MMI≥VIII) from a regional fault source. Seismic shaking from the historical rupture of the eastern Awatere Fault during the M 7.5 1848 Marlborough earthquake is considered to be the best candidate for such a change in the sediment flux coming from Goat Valley Stream. Other regional historical shaking events such as the 1855 Wairarapa, 1888 Amuri and 1929 Buller earthquakes may have been distant or weak enough to have not generated increased sediment flux or failure within the lake.

Our reconnaissance study of Lake McRae indicates that it has significant potential for further landslide, paleoearthquake, ground motion, and lacustrine studies and also has potential as a site for paleoclimate study. Future research at the lake could include intermediate (6-m) depth coring of sediments, further cosmogenic dating of the landslides and study of the activity and paleoseismicity  of the Clarence Fault. In light of the recent November 14, 2016 Mw 7.8 Kaikoura earthquake sequence, further work on the active faulting and sedimentary record of this important small basin is vital. The effects of shaking and turbidity related to the Kaikoura earthquake should be a tractable target of future research.